Jump to content

Digital Signal Processing/Impulse Response

From Wikibooks, open books for an open world

Let's say that we have the following block diagram:

h[n] is known as the 'Impulse Response of the digital system. We will talk about impulse responses here in this chapter.

Impulse Function

[edit | edit source]

In digital circuits, we use a variant of the continuous-time delta function. This delta function (δ[n]) has a value of 1 at n = 0, and a value of 0 everywhere else.

δ[n] = [... 0 0 0 0 1& 0 0 0 0 ...]

If we set x[n] = δ[n], then the output y[n] will be the response of the system to an impulse, or simply, the "Impulse Response".

We can time-shift the impulse function as such:

δ[n-1] = [... 0 0 0 0& 1 0  0 0 0 ...]
δ[n+1] = [... 0 0 0 0  1 0& 0 0 0 ...]

We can add time-shifted values of the impulse function to create an Impulse Train:

y[n] = δ[n] + δ[n-1] + δ[n-2] + [n-4]
y[n] = [1& 1 1 0 1]

Impulse Trains

[edit | edit source]

Impulse Response

[edit | edit source]

If we have a difference equation relating y[n] to x[n], we can find the impulse response difference equation by replacing every y with an h, and every x with a δ:

y[n] = 2x[n] + 3x[n-1] + 4x[n-2]
h[n] = 2δ[n] + 3δ[n-1] + 4δ[n-2]

And by plugging in successive values for n, we can calculate the impulse response to be:

h[n] = [2& 3 4]

Output

[edit | edit source]

Now, let's say that we have a given impulse response, h[n], and we have a given input x[n] as such:

x[n] = [1& 0 1 2]
h[n] = [2& 2 2 1]

We can calculate the output, y[n], as the convolution of those 2 quantities:

x[n]    ->       1& 0 1 2
h[-n]   ->       1  2 2 2&
h[-n-3] -> 1 2 2 2&               -> y[m-3] = 2&
h[-n-2] ->   1 2 2  2&            -> y[m-2] = 2
h[-n-1] ->     1 2  2 2&          -> y[m-1] = 4
h[-n]   ->       1  2 2 2&        -> y[m]   = 7
h[-n+1] ->          1 2 2 2&      -> y[m+1] = 6
h[-n+2] ->            1 2 2 2&    -> y[m+2] = 5
h[-n+3] ->              1 2 2 2&  -> y[m+3] = 2

y[n] = [2& 2 4 7 6 5 2]